Rapid identification and phylogenetic classification of diverse bacterial pathogens in a multiplexed hybridization assay targeting ribosomal RNA

Roby P Bhattacharyya,Jamin Liu,Maura Benson,Sharon Y Wong,Peijun Ma,Noam Shoresh,Mark Walker,Jennifer Skerry,Austin C Hachey,Marcia B Goldberg,Rich Boykin,Joseph M Beechem ,Virginia Pierce ,Kaelyn C Cummins,Hyunryul Ryu,Jongyoon Han,Sophie S Son,Kyungyong Choi,Lisa A Cosimi,Lidan Wu,Jonathan Livny,Deborah T Hung

doi:10.1038/s41598-019-40792-3

Abstract

Rapid bacterial identification remains a critical challenge in infectious disease diagnostics. We developed a novel molecular approach to detect and identify a wide diversity of bacterial pathogens in a single, simple assay, exploiting the conservation, abundance, and rich phylogenetic content of ribosomal RNA in a rapid fluorescent hybridization assay that requires no amplification or enzymology. Of 117 isolates from 64 species across 4 phyla, this assay identified bacteria with >89% accuracy at the species level and 100% accuracy at the family level, enabling all critical clinical distinctions. In pilot studies on primary clinical specimens, including sputum, blood cultures, and pus, bacteria from 5 different phyla were identified.

Highlights

After long relying on decades-old culture methods and traditional biochemical assays, modern clinical microbiology laboratories have begun to incorporate novel methods for bacterial pathogen identification in recent years[1]
We previously demonstrated that NanoString assays can identify a limited set of bacterial Ribosomal RNA (rRNA) targets with high sensitivity[33]; here, we sought to take on the challenge of designing a probeset that would span the breadth and complexity of phylogenetic diversity encompassed by clinically relevant bacterial pathogens, thereby detecting and identifying a wide range of bacterial species in a single, simple assay
Unlike typical quantitative hybridization-based assays such as microarrays or mRNA-directed NanoString assays in which signal intensity reflects abundance of the target, here, signal intensity instead reflects the hybridization efficiency of the probe pair for different targets, governed by the degree of sequence complementarity between each probe pair and its target. (This assumes that the 16S and 23S rRNA targets are in fixed relative abundance.) Bacterial identity is inferred from the ensemble pattern of reactivity across the 180 probe pairs with the target bacterial rRNA

Summary

Introduction

After long relying on decades-old culture methods and traditional biochemical assays, modern clinical microbiology laboratories have begun to incorporate novel methods for bacterial pathogen identification in recent years[1]. RRNAs are by far the most abundant RNA species in a bacterium; at >10,000 copies per cell[25,26,27], they typically comprise >85% of the RNA mass of a cell[28,29] and serve as an abundant target for direct characterization even in the absence of amplification Taking advantage of these features of rRNA, we designed and developed an approach for broad-range bacterial identification through multiplexed hybridization of fluorescent DNA probes to ribosomal RNA, with the goal of matching the advantages of unbiased sequence-based pathogen identification in a simplified, more rapid applied genomics assay that requires no amplification or enzymology. This nuanced interpretation of hybridization efficiency as a proxy for sequence relies on the accurate, quantitative nature and broad linear dynamic range of the NanoString assay platform, as well as its ability to multiplex across hundreds of targets in a single reaction[30], offering advantages over prior efforts at rRNA-directed, hybridization-based pathogen identification using microarrays[35,36,37] or fluorescence in situ hybridization[25]

Objectives

Methods

Results

Conclusion